WO1992002553A1

WO1992002553A1 - Tumor cell binding molecule with protease inhibiting region

Info

Publication number: WO1992002553A1
Application number: PCT/GB1991/001322
Authority: WO
Inventors: David James Ballance; Michael George Courtney
Original assignee: Delta Biotechnology Limited
Priority date: 1990-08-03
Filing date: 1991-08-02
Publication date: 1992-02-20
Also published as: AU8318591A; GB2246779A; GB9017083D0; GB2246779B

Abstract

A molecule comprising a first region which binds to a tumour (preferably the receptor for uPA) and a second region which inhibits a (preferably tumour-associated) protease, for example uPA. The first and second regions may be respective antibodies or parts thereof specific for appropriate parts of uPAR and uPA. Preferably, the first region is a uPAR-binding part of uPA, e.g. the 12-32 region theeof, and the second region is PAI-2 or a uPA-inhibiting analogue or part thereof, and the first and second regions are combined to form a single polypeptide, expressible from a transformed host.

Description

TUMOR CELL BINDING MOLECULE WITH PROTEASE INHIBITING REGION

The process of activation of the zymogen plasminogen to the broad spectrum serine protease, plasmin, is mediated by specific plasminogen activators: tissue-type plasminogen activator (tPA) and urokinase-type plasminogen activator (uPA) . tPA is primarily responsible for the generation of plasmin in fibrinolysis and its activity is stimulated when bound to fibrin. In humans, and many other species, uPA is important in generating proteolytic activity responsible for extracellular matrix and basement membrane degradation associated with tissue remodelling. Such processes are involved in growth and spread of tumour cells, mammary gland involution, ovulation, embryo development, development of the nervous system, in the normal inflammatory response and in a number of inflammatory diseases (Danø et al , 1985) . Elevated levels of uPA have been found in malignant tissue and there is believed to be a general correlation between the amount of uPA and the invasiveness of the tumour. Studies of experimental Lewis lung tumours revealed concentration of uPA at the invasive edge of the tumours. In addition, anti-catalytic antibodies to uPA inhibited the establishment of tumour cells after i.v. injection into mice and inhibited metastasis in chick embryos. The proteolytic activity released by uPA is localized by the binding of plasminogen, plasmin and uPA to specific cell surface receptors (Vassalli et al, 1985; Miles and Plow, 1988) and the importance of receptor-bound uPA in tissue invasion has been demonstrated (Ossowski, 1988). The receptor-binding region has been localized to the growth factor domain of the uPA molecule in the region of amino acids 12-32 of mature uPA since peptides corresponding to this sequence are able to block the uPA receptor interaction by competition (Appella et al , 1987).

The apparent central role of uPA in malignant disease indicates that inhibition of the activity of uPA might modify the course of the disease. Of the natural inhibitors of uPA, only plasminogen activator inhibitor 1 (PAI-1) and plasminogen activator inhibitor 2 (PAI-2) are sufficiently active and specific enough to be considered (Sprengers and Kluft, 1987) . Although PAI-1 is a faster inhibitor of uPA than PAI-2 it is also a very effective inhibitor of tPA and administration of this protein would be likely to compromise fibrinolysis. PAI-2, on the other hand, is a slower inhibitor of tPA and is reported to be inactive against the physiologically important fibrin- stimulated tPA. PAI-2 has been shown to inhibit receptor-bound uPA (Kirchheimer and Remold, 1989a; Pδllanen et al, 1990) and uPA-dependent tissue invasion (Kirchheimer and Remold, 1989b) and thus represents a potentially useful anti-tumour or anti- metastatic agent. Natural sources of PAI-2 yield very small quantities of the protein but EP-A-238 275 discloses the production of recombinant PAI-2 in E. coli . Our co-pending application GB8918191.1 discloses the advantageous production of recombinant PAI-2 in Saccharomyces cerevisiae .

Fusions of PAI-2 and other polypeptides are described in EP- A-238 275 (page 30) but only in the context of obtaining secretion of the PAI-2 from a transformed cell producing it or in the context of detecting PAI-2 production as a β-galactosidase fusion.

WO 88/08451 discloses uPA-tPA fusions, but the portion of uPA which is used is the B chain and the intention is to retain the activities of both tPA and uPA. As is described below, the uPA-derived portion (if present) of the molecules of the present invention preferably has substantially no uPA-like a idolytic or proteolytic activity.

One aspect of the invention provides a molecule comprising a first region which binds to a tumour cell and a second region which inhibits a first protease.

The specific tumour-binding may be enabled by the existence on the tumour of structures unique to the tumour, for example tumour-specific antigens such as CEA (carcino-embryonic antigen) , pan carcinoma antigen, placental alkaline antigen or polymorphic epithelial mucin antigen. Alternatively, the specificity may arise from the existence of a higher level or density on the tumour cell surface of a structure which is found on normal cells; the receptor for uPA (uPAR) is an example of such a structure.

The tumour-binding region is preferably an entity which will bind to the receptor for uPA (uPAR) , as is described in more detail below. Alternatively, it may be any other entity which binds preferentially to a tumour cell, for example the epidermal growth factor receptor present in some breast carcinomas.

The first protease is preferably tumour-associated and is preferably urokinase-type plasminogen activator (uPA) as is described in more detail below. However, it could be any other protease of which high local levels are associated with tumours, or at least tumours of a given kind. Preferably the protease is one that can be activated or converted from a pro-enzyme -or zymogen form by uPA or receptor-bound uPA or by plasmin to generate localised proteolytic activity. The inhibitor region in the molecule of the invention may be an inhibitor of plasmin such as the Pittsburgh variant of α^-antitrypsin, an α_χAT variant having lysine at the P position, α₂-antiplasmin, α₂- macroglobulin, aprotinin or any other inhibitor with plasmin inhibitory activity, or it may be an inhibitor of a collagenase such as tissue inhibitor of metallo-proteinases (TIMP) or the related TIMP2. Other proteases believed to be involved in tissue remodelling in neoplasia are proteases of the cathepsin and stromelysin families and it may be beneficial to provide an inhibitor of these activities. Another protease (not tumour- associated) is thrombin, a potent activator of platelet aggregation which activates a number of other cell types, including endothelial cells, thus initiating their response to stress. This serine protease is consequently an important regulator of thrombosis and haemostasis in addition to its role in generation of fibrin. The action of thrombin on platelets and endothelial cells is mediated by the thrombin receptor (Vu et al , 1991) and activation of this receptor involves the cleavage of an extracellular domain of the receptor by thrombin. Thus, the activation of the thrombin receptor and the consequent activation of platelets and other cell types may be inhibited by a thrombin inhibitor such as α₁-antitrypsin Pittsburgh. An application of the concept of the current invention to this system is to enhance the anti-thrombin-receptor-activating activity of thrombin by directing the α_χ-AT Pittsburgh to the thrombin receptor or an adjacent cell surface molecule. In this instance the receptor- binding moiety of the hybrid may be a receptor-binding peptide or an antibody or antibody fragment directed against the thrombin receptor. In all of these cases an inhibitory portion of the protease inhibitor may be used instead of the whole molecule.

The molecule thus provides for the targeting of the protease inhibiting region to the tumour cell.

.The protease the activity of which is inhibited may be free, bound to its receptor, associated with the cell surface via non¬ specific interactions or bound to adjacent cells or to extracellular matrix.

The first region is preferably a peptide sequence corresponding to a receptor-binding a region of the tumour- associated protease, which may or may not be the same protease as the protease which is inhibited. Thus, the first region may comprise amino acids 1-32 of uPA or a variant or fragment thereof. By "variant" we mean a region with one or more minor alterations to the amino acid sequence thereof which region is nevertheless recognisably similar to the said 1-32 region and which still binds to the uPA-binding site on the uPAR. Incidentally, although the region is generally "EGF-like", epidermal growth factor (EGF) itself does not bind to the said site. By "fragment" we mean smaller sequences (preferably at least 10, 15, 19 or 20 amino acids long) which still bind to the binding site of the receptor, for example the 20-30 region of uPA. "Variants" of "fragments" may also be used. The receptor- binding region of the protease may, in the molecule of the invention, be accompanied by other parts of the protease or a pro-form of the protease or a mutant thereof. Such parts or mutants should have at least 50% homology, preferably 60%, 70%, 80%, 90% or 95% homology with the protease (or, in the case of parts, with the corresponding region of the protease) in order to be regarded as parts or mutants thereof. They should, however, at least when present in the molecule of the invention, not have any undesirable level of protease-like proteolytic activity, since otherwise the point of the invention would be lost. Preferably, the protease-derived part of the molecule of the invention, at least when present in the said molecule, has substantially no protease-like proteolytic activity. In the case of uPA, it may be possible to achieve this simply by using a pro- uPA mutant in which Lys¹⁵⁸ has been altered, for example to Gin or Gly, to prevent cleavage by plasmin. The 12-19 and 31-32 residue regions adjacent the 20-30 bindir.g region of uPA seem to help the binding region to bind optimally and are preferably present, or functionally equivalent flanking regions are present. The first region may alternatively be a monoclonal antibody or part thereof or a genetically engineered counterpart thereof which binds to the protease receptor, preferably uPAR. Preferably, it binds to the protease-binding site of the receptor or it binds to an area of the receptor which is sufficiently close to the said binding site for the bulk of the molecule as a whole to block binding of the protease to the receptor. In this way, not only is the protease-inhibiting region targeted to the receptor, but the protease-binding site of the receptor is blocked, which will prevent binding of the protease to that receptor and thus may further help to prevent the protease from acting on its substrate. Monoclonal antibodies to the protease receptor or other target entity may be prepared by immunising an animal with the target entity or an antigenic part thereof and preparing and selecting an appropriate hybridoma in the conventional way (see Monoclonal Hybridoma Antibodies Ed. J.G.R. Hurrell, CRC Press 1982, incorporated herein by reference). Alternatively, a polyclonal collection of antibodies of restricted specificity may be used. Preferably, however, the antibody is a monoclonal antibody, an F_ab', (F_ab)₂, F_v, scF_v or dAb fragment or a synthetic molecule including the minimum recognition unit(s) of a target-entity-specific immunoglobulin.

In other embodiments, the first region consists of any peptide which will bind to the protease-binding site, or sufficiently close to it to block it, and need not be homoxogous with any part of the protease. The second region may be any compound which inhibits the action of the protease (preferably uPA) . Such a compound may be an antibody (or antibody fragment etc, as discussed above) directed against the enzymatically active site of the protease, or against such a part of the protease molecule as to block access of the substrate to the active site or against any other part of the protease such that binding of the antibody to the said part prevents or at least reduces the activity of the protease towards its substrate.

In a particularly preferred embodiment, however, the second region is PAI-2 or part thereof or a mutant PAI-2 or part thereof. Preferably, the mutant or part of PAI-2 retains a useful level of the uPA-inhibiting activity of PAI-2. Fusions of the uPAR-binding domain of uPA and PAI-2 or a part or mutant thereof (hereinafter called "uPA-PAI-2 fusions") may exhibit the following properties:

1. binding unoccupied uPAR (ie uPAR unoccupied by uPA) via the growth factor domain, thereby blocking binding by uPA to uPAR,

2. in solution, inhibiting free or receptor-bound uPA, and

3. when receptor-bound, inhibiting receptor-bound uPA.

The molecule should be stable In vivo. By "stable in vivo" , we mean that the molecule is sufficiently stable when administered parenterally, preferably intravenously, to allow useful quantities of it to reach a tumour and bind to the receptors on the tumour cells. In practice, a molecule with a half-life in the human circulation of 30 minutes or longer is considered to be stable. Preferably, the half-life is at least 1 hour, 24 hours, 3 days or more.

The molecules of the invention may be dissolved in suitable delivery vehicles and administered to patients to inhibit or prevent the growth or spread of an actual or suspected tumour. The patient is preferably a human but may be another animal, preferably a mammal, such as a pet (dogs, cats, etc) or an economically important animal (sheep, cattle, pigs, fowl, horses, etc) . In the case of administration to animals, the tumour- binding domain is modified for optimal binding to the animal receptor and ideally is the sequence of the said domain of the protease of that animal. In addition, the inhibitor moiety may be modified to effect efficient inhibition of the protease of that animal. The molecules are preferably administered parenterally, for example intravenously, intramuscularly or subcutaneously, by injection or infusion. Clinically qualified people will be able to determine suitable dosages, delivery vehicles and administrative routes.

The molecules of the invention may be produced by chemically linking the said first and second regions by methods known to those in the art of protein chemistry, for example using the methods of O¹Sullivan et al (1979) and bifunctional linking reagents suchasm-maleimido-benzoyl-N-hydroxy-succinimideester. However, it is preferred for the molecules to be produced by expression of a recombinant DNA sequence in a suitable host transformed therewith. Such hosts are legion and include E . coli , B . subtil is, Aspergillus and other filamentous fungi, yeasts, animal cells, plant cells, insect cells and transgenic whole animals and plants.

Exemplary genera of yeast contemplated to be useful in the practice of the present invention are Pichia, Saccharomyces , Kl uyveromyces , Candi da , Torulopsis , Hansenul a ,

Schizosaccharomyces, Citeromyces, Pachysolen, Debaromyces, Metschunikowia, Rhodosporidium, Leucosporidium, Botryoascus , Sporidiobolus , Endomycopsis , and the like. Preferred genera are those selected from the group consisting of Pichia, Saccharomyces, Kluyveromyces, Yarrowia and Hansenula , because the ability to manipulate the DNA of these yeasts has, at present, been more highly developed than for the other genera mentioned above.

Examples of Saccharomyces are Saccharomyces cerevisiae, Saccharomyces italicus and Saccharomyces rouxii .

Examples of Kluyveromyces are Kluyveromyces fragilis and Kluyveromyces lactis .

Examples of Hansenula are Hansenula polymorpha, Hansenula anomala and Hansenula capsulata . Yarrowia lipolytica is an example of a suitable Yarrowia species.

Yeast cells can be transformed by: (a) digestion of the cell walls to produce spheroplasts; (b) mixing the spheroplasts with transforming DNA (derived from a variety of sources and containing both native and non-native DNA sequences) ; and (c) regenerating the transformed cells. The regenerated cells are then screened for the incorporation of the transforming DNA.

It has been demonstrated that yeast cells of the genera Pichia, Saccharomyces , Kluyveromyces , Yarrowia and Hansenula can be transformed by enzymatic digestion of the cell walls to give spheroplasts; the spheroplasts are then mixed with the transforming DNA and incubated in the presence of calcium ions and polyethylene glycol, then transformed spheroplasts are regenerated in regeneration medium.

Methods for the transformation of S. cerevisiae are taught generally in EP 251744, EP 258067 and WO 90/01063, all of which are incorporated herein by reference.

In the case of PAI-2-based fusions, the protein is preferably produced as a soluble intracellular protein.

Suitable promoters for S. cerevisiae include those associated with the PGK1 gene, GAL1 or GAL10 genes, CYC1, PH05, TRP1 , ADH1, ADH2, the genes for glyceraldehyde-3-phosphate dehydrogenase, hexokinase, pyruvate decarboxylase, phosphofructokinase, triose phosphate isomerase, phosphoglucose isomerase, glucokinase, α-mating factor pheromone, a-mating factor pheromone, the PRB1 promoter, the GOT2 promoter, and hybrid promoters involving hybrids of parts of 5' regulatory regions with parts of 5• regulatory regions of other promoters or with upstream activation sites (eg the promoter of EPA-258067) . The preferred promoter is the PRB1 promoter.

The transcription termination signal is preferably the 3 ' flanking sequence of a eukaryotic gene which contains proper signals for transcription termination and polyadenylation. Suitable 3¹ flanking sequences may, for example, be those of the gene naturally linked to the expression control sequence used, ie may correspond to the promoter. Alternatively, they may be different in which case the termination signal of the S. cerevisiae ADH1 gene is preferred.

Suitable secretion leader sequences, if the molecule is to be secreted from the host, include mammalian leader sequences, such as the HSA and pro-uPA leader sequences, S . cerevisiae leader sequences such as the α-mating factor pheromone pre- and prepro- sequence, the invertase (SUC2 ) leader sequence, the PH05 leader sequence, or hybrid leader sequences such as the leader sequence of WO 90/01063.

The amino acid sequence of the N-terminal portion of uPA encompassing the growth factor domain is as follows: 10 20 30

S N E L H Q V P S N C D C L N G G T C V S N K Y F S N I H W C N

40 50

C P K K F G G Q H C E I D K S K T C

The sequence of residues 20-30 is believed to be responsible for the specificity of binding to the receptor and this sequence may represent the minimum requirement for binding to the receptor. Sequences adjacent to this region provide the proper conformation for optimum binding but may be substituted by a functionally equivalent structure. The uPA sequence incorporated into the fusion molecules preferably includes residues 12-32 and in the preferred embodiment includes residues 1-47 to ensure optimum binding to uPAR. It is not necessary for the first region of the molecule of the invention to be at or adjacent the N-terminal of the molecule. The second region can be N-terminal to the first region.

DNA sequences for expression of the molecules of the invention may be prepared by known techniques, for example by fusing cDNA fragments prepared from mRNAs corresponding respectively to the receptor-binding region of the protease and to a protease inhibitor or a part thereof.

Monoclonal antibodies may be prepared generally by the techniques of Zola (1988) which is incorporated herein by reference. Useful antigens for preparing either monoclonal or polyclonal antibodies are (for the second region) uPA, low molecular weight uPA or the heavy (B) chain of uPA, and (for the first region) uPAR or peptides based on the uPA-binding domain thereof. Antibody fragments such as F_ab fragments, may be prepared therefrom in known ways. The antibodies may be humanized in known ways. Antibody-like molecules may be prepared using the recombinant DNA techniques of WO 84/03712.

The art of "antibody engineering" is advancing rapidly, as is described in Tan & Morrison (1988) , Williams (1988) and Neuberger et al (1988) (all of which are incorporated herein by reference) , and is well suited to preparing the first or second regions of the molecules of the invention or the whole molecule itself. Thus, bispecific antibodies, specific for an appropriate receptor and an appropriate site on the protease, may be made by any of the methods described by Williams (1988) . Such molecules will block the protease-binding site of the receptor and also inhibit the proteas^'e. The -antibody may alternatively be bispecific for the receptor and the protease inhibitor (or a mutant or fragment thereof) so that the antibody acts to bind the inhibitor to the receptor. The combination of the antibody bound to the inhibitor (or a mutant or fragment thereof) thus constitutes a molecule of the invention, whereas the antibody itself constitutes a precursor thereof, forming a separate aspect of the invention. The antibody specific for the receptor and the inhibitor may be combined with the inhibitor before being administered or may be co-administered therewith. By "co- administered", we do not necessarily mean that the antibody and the inhibitor must be administered simultaneously; it is sufficient for them to combine in a therapeutically useful way in the body.

Chimaeric antibodies, where the F_c region is a human immunoglobin or part thereof, may be desirable for long term treatment, to reduce adverse im unological responses. Single chain antibodies may be used for either or both of the two regions of the molecules of the invention.

Preferred embodiments will now be described by way of example and with reference to the accompanying figures.

Figure 1 illustrates the construction of plasmids pDBPl, pDBP2 and pDBP3. E = EcoRI , P = PstI, Bg = Bglll, H = tfindlll, B = BajjiHI and A = Aflll. Only the Alll site present in the PAI-2 encoding sequence is shown;

Figure 2 (on four sheets) shows a DNA sequence encoding, and corresponding amino acid sequence of, PAI-2 in pDBPl;

Figure 3 shows the DNA sequence of the PRB1 promoter;

Figures 4 to 8 respectively illustrate plasmids pAYE333, PAYE334, pAYE335, pDBP5 and pDBP6;

figure 9 illustrates the construction of plasmids pHJB117 and pDJBllδ; Figure 10 illustrates the construction of plas id pDBUPl;

Figure 11 shows flow cytometer histograms of U937 cells to detect binding of the uPA-PAI-2 hybrid protein to cells. Control cells are shown by the left hand peak and cells incubated with the hybrid protein are shown by the right hand peak. Detection of bound hybrid protein was with anti-PAI-2 antibody and FITC- conjugated secondary antibody;

Figure 12 illustrates the construction of plasmids pDBA2, pDBUAl, pDBUA2 and pDBUA3.

Figure 13 shows the DNA sequences and amino acid sequence encoded thereby, for o^antitrypsin in pDBAl.

Example 1

Standard recombinant DNA procedures are as described by Sambrook et al, 1989 unless otherwise stated.

General

A plasminogen activator inhibitor 2 coding sequence (Figure 1) may be derived from EP-A-238 275. The DNA may be obtained from E. coli deposited as ATCC 53585 in connection with EP 238 275. This contains the plasmid pBTA438, comprising the PAI-2 sequence. The sequences encoding the uPA growth factor domain may be assembled from synthetic oligonucleotides by standard procedures.

However, we obtained a cDNA sequence for PAI-2 ourselves by analogous techniques. An expression vector was constructed in which DNA encoding amino acids 1-47 of uPA, preceded by a methionine initiation codon, directly preceded DNA encoding PAI-2 such that the two coding sequences were in frame. Expression of these sequences was under the control of the S . cerevisiae PRB1 gene promoter and transcription termination was effected by the S . cerevisiae ADH1 terminator.

Specific Detail

PAI-2 Coding Sequence

A lambda gtll cDNA library constructed from RNA isolated from 4-phorbol-12-myristate-13-acetate stimulated cells of the human monocyte-like histiocytic lymphoma cell line U937 (obtained from Clontech Laboratories Inc) was used as a source of PAI-2 cDNA. The library was screened using radioactively labelled oligonucleotide probes corresponding to the DNA sequences encoding the N-terminus (oligo 1) and C-terminal end (amino acids 400-410) (oligo 2) of the PAI-2 protein, respectively by standard procedures.

Oligo 1

5*-ATG GAG GAT CTT TGT GTG GCA AAC ACA CTC TTT-3 Oligo 2

5 ^»-GCC GAA AAA TAA AAT GCA CTT GGT TAT CTT ATG-3 ^»

From the putative positive clones we selected one clone (lambda gtll-186) which appeared to contain the entire PAI-2 coding region. This was confirmed by sequence analysis of the DNA insert in this clone following transfer to M13mpl9 to form pDBPl (Figure 1) . The sequence is shown in Figure 2.

To facilitate insertion into expression vectors, restriction enzyme recognition sites were created at the 5' and 3¹ ends of the PAI-2 gene. A Bglll site was created at the 5¹ end of the gene using the anti-sense oligonucleotide primer 1:

5 '-TGCCACACAAAGATCTTCCATTGTTTCAATCT-3 • to create a mutation in the third position of the second codon as shown below:

M E D L C V A 5 ' ...AGATTGAAACA ATG GAG GAT CTT TGT GTG GCA...3 • 3 ' ...TCTAACTTTGT TAC CTC CTA GAA ACA CAC CGT...5' changed to:-

M E D L C V A 5 ' ...AGATTGAAACA ATG GAA GAT CTT TGT GTG GCA...3 • 3 ' ...TCTAACTTTGT TAC CTT CTA GAA ACA CAC GCT...5 '

Bglll

An A-flll site was created at the 3 ^« end of the gene using the anti-sense oligonucleotide primer 2:

5•-CAGAAGCAGCACGCTTAGTCTTAAGGTGAGGAAATCTGCC-3 ' to create mutations in the third position of the last codon (proline) and in the first base after the stop codon as shown below:-

G R F S S P STOP 5• ...GGC AGA TTT TCC TCA CCC TAA AACTAAGCGTGCTGCTTCTG...3 ' 3 ' ...CCG TCT AAA AGG AGT GGG ATT TTGATTCGCACGACGAAGAC...5' changed to:

G R F S S P STOP 5' ...GGC AGA TTT TCC TCA CCT TAA GACTAAGCGTGCTGCTTCTG...3 '

3 '...CCG TCT AAA AGG AGT GGA ATT CTGATTCGCACGACGAAGAC...5'

I i

Aflll

These two oligonucleotides were annealed to single stranded pDBPl and then used in an in vitro mutagenesis procedure (oligonucleotide-directed in vitro mutagenesis system-version 2, Amersham pic) carried out according to the manufacturer's recommendations. A clone derived from this procedure and with the correct changes was designated pDBP2 (Figure 1) .

Oligonucleotide linkers were then used to position restriction sites at either end of the gene which are suitable for insertion of the gene into an expression vector. The linker positioned at the 5' end of the gene was Linker 1

5'-AGCTTGTCGACGGATCCAAAAAG ATG GAA

ACAGCTGCCTAGGTTTTTC TAC CTT CTAG-5'

I I I I I I

HindiII BamHI Bgrlll and the 3 ' linker was Linker 2

flindlll

I 1 5•-TTAAGTCGACAAGCTTG

CAGCTGTTCGAACCTAG-5'

1 I I I

Aflll BamKI

These two linkers were ligated with the Bglll-Aflll PAI-2 gene fragment from pDBP2 into Hindlll + BamHI digested M13mpl9 to form pDBP3 (Fig 1) .

PRB1 Promoter

The structural gene, PRB1, for the Saccharomyces cerevisiae vacuolar endoprotease B has been isolated- (Moehle et al, 1987a) on two prJbl complementing plasmids called MK4 and FP8. When the yeast Saccharomyces cerevisiae is grown on glucose in shake flask culture, very little protease B activity is detected until the cells have catabolised the glucose and are utilising the ethanol accumulated during growth (Saheki and Holzer, 1975; Jones et al, 1986) . This is believed to be a consequence of a transcriptional control mechanism which represses mRNA accumulation until the glucose has been exhausted and the culture enters the diauxic plateau (Moehle et al , 1987a). Studies with protease B (prbl~) deficient mutants implicate protease B in the protein degradation that occurs when negative cells are starved of nitrogen and carbon (Wolf and Ehmann, 1979; Zubenko and Jones, 1981).

The DNA sequence of the PRB1 gene has been reported, as has 150bp of the PRB1 promoter (Moehle et al, 1987b). A more extensive DNA sequence of the PRB1 promoter is also available as an entry in the Genbank database, release 60, accession number M18097, locus YSCPRB1, Figure 3.

The whole of the PRB1 promoter may be used, or a smaller portion thereof, as may readily be determined. For example, the roughly lkbp sequence extending upstream from the start codon to the SnaBl site is effective.

The 1.435kbp #indIII-_.coRI DNA fragment containing the protease B promoter (Figure 3) was cloned into the polylinker of the M13 bacteriophage mpl8 (Yanisch-Perron et al , 1985), generating plasmid pAYE333 (Figure 4) . Double stranded pAYE333 was linearised by partial digestion with SnaBl and the double stranded oligonucleotide linker 3 inserted by ligation. Linker 3

5'-GCGGCCGC-3 '

3 ^»-CGCCGGCG-5^•

I I

WotI

This generated a WotI restriction site at the 5¹ end of the protease B promoter. The promoter element was further modified by site directed mutagenesis (oligonucleotide direct in vitro mutagenesis system-Version 2, Amersham) according to the manufacturer's instructions. Mutagenesis with the oligonucleotide

5 '-CGCCAATAAAAAAACAAGCTTAACCTAATTC-3 '

introduced a Hindlll restriction site close to the ATG translation initiation codon:

CGCCAATAAAAAAACAAACTAAACCTAATTCTAACAAGCAAAGATG

I I Met

CGCCAATAAAAAAACAAGCTTAACCTAATTCTAACAAGCAAAGATG

^{1 1} Met

HindiII Plasmid pAAH5 (Goodey et al, 1987) was linearised by partially digesting with BamHI. The 5' protruding ends were blunt-ended with T4 DNA polymerase and ligated with the double- stranded oligonucleotide Linker 3. A recombinant plasmid pAYE334 (Figure 5) was selected in which a WotI restriction site had replaced the BamHI site at the 3' end of the ADH1 terminator.

The 0.8kbp Wotl-Hindlll modified protease B promoter sequence was placed upstream of the 0.45kbp tfindlll-Wotl ADH1 transcription terminator on a pAT153-based plasmid (Twigg and Sherratt, 1980) to generate pAYE335 (Figure 6) .

The large 6.38kbp Hindlll-BamHI fragment from the yeast E. coli shuttle vector pJDB207 (Beggs, 1981) was treated with the Klenow fragment of E . coli DNA polymerase to create flush ends and ligated with the double stranded olignonucleotide Linker 3 to generate plasmid pDBP5 (Figure 7) .

The 1.25 kbp WotI Protease B promoter /ADH1 terminator cassette from plasmid pAYE335 (Figure 6) was introduced into the unique WotI site of plasmid pDBP5, generating pDBP6 (Figure 8) .

uPA-PAI-2 fused coding region

Two oligonucleotides were synthesised using an Applied Biosystems Inc 380B oligonucleotide synthesiser and annealed to form a linker (linker 5) to encode the C-terminal portion of the uPA growth factor domain (amino acids 24-47 of mature uPA) and the start of the PAI-2 sequence. Linker 5

Y F S N I H W C N C P K K F G TCGAGTAC TTC TCC AAC ATT CAC TGG TGC AAC TGC CCA AAG AAA TTC GGA CATG AAG AGG TTC TAA GTG ACC ACG TTG ACG GGT TTC TTT AAG CCT

I I

Seal

G Q H C E I D K S M E GGG CAG CAC TGT GAA ATA GAT AAG TCA ATG GAA

CCC GTC GTG ACA CTT TAT CTA TTC AGT TAC CTT CTA G

I i

Bglll

This linker was ligated with a Bglll-Bindlll fragment of pDBP3, representing the remainder of the PAI-2 sequence, into pUC18 at Sall-Hindlll to form the plasmid pDJB117 (Figure 9) .

A second linker representing the PRB1 ATG environment, a methionine initiation codon and DNA encoding the N-terminal portion of the uPA growth domain (amino acids 1-23 of mature uPA) was assembled by annealing the four oligonucleotides shown below:- Linker 6 oligo 3

M S N E L H Q V GATCAAGCTTAACCTAATTCTAACAAGCAAAG ATG AGC AAT GAA CTT CAT CAG GTA

TTCGAATTGGATTAAGATTGTTCGTTTC TAC TCG TTA CTT GAA GTA GTC CAT

I i

HindiII i

oligo 6

oligo 4

P S N C D C L N G G T C V S N K CCA TCG AAC TGT GAC TGT CTA AAT GGA GGA ACA TGT GTG TCC AAC AAG GGT AGC TTG ACA CTG ACA GAT TTA CCT CCT TGT ACA CAC AGG TTG TTC

oligo 5

T A

The plasmid pDJB117 was partially digested with Seal and linearised DNA was isolated from an agarose gel and then digested with BamHI. This DNA was ligated with linker 6 to form pDJBllδ (Figure 9) . The ffindlll fragment from pDJBllδ (Figure 9) containing the uPA-PAI-2 coding sequence was then introduced into the expression plasmid pDBP6 to form pDBUPl (Figure 10) . Plasmid pDBUPl contains a PRBl promoter fragment to control transcription of the uPA-PAI-2 sequence and the ADH1 transcription terminator. It also contains sequences of the bacterial plasmid pAT153 (Twigg and Sherratt, 1980) , the leu2d gene for selection of transformants and part of the 2μm plasmid to provide replication function in S. cerevisiae . The plasmid was introduced into a suitable S. cerevisiae strain wherein it effected production of the hybrid protein as a soluble intracellular protein which was recognised by anti-PAI-2 and anti-uPA antibodies in Western blotting experiments.

The protein was purified from cell lysates by copper chelate and anion exchange chromatography and was found to inhibit uPA with similar kinetics to PAI-2 itself, ie second order inhibition rate constant of l.δ x 10⁶ M^'^-S^"1, determined according to Thorsen et al (1988) .

The uPA growth factor domain was shown to be functional by its ability to mediate specific binding to U937 cells. The cells were first treated with 50mM glycine HCl, 0.1M NaCl, pH3.0 for 3 minutes at 4°C to remove endogenous cell-bound uPA. The cell suspension was neutralised by the addition of one-fifth volume of 0.5M Hepes, 0.1M NaCl, pH7.5 and then the cells were washed twice with PBS (Dulbecco's Ca²⁺ and Mg²⁺ free phosphate buffered saline) containing 0.1% (w/v) bovine serum albumin (PBS + BSA). Aliquots of cells (0.5 x 10⁶) were then incubated with or without approximately 0.5μg uPA-PAI-2 hybrid protein in PBS + BSA for 90 min on ice and then for a further 60 min with 10% (v/v) pooled normal human serum in Dulbecco's PBS with 20mM glucose, 5mM Na₂EDTA, 0.1% (w/v) BSA, 0.1% (w/v) NaN₃ (PBS + NHS) to block cell membrane F_c receptors. The cells were washed three times with PBS + NHS and then incubated with goat anti-PAI-2 antibody (American Diagnostica Inc) in PBS + NHS for 30 min on ice. The cells were then washed twice with PBS + NHS and then incubated with anti-goat IgG antibody conjugated to fluorescein isothiocyanate in PBS + NHS for 30 min on ice. Finally the cells were washed twice with PBS + NHS and resuspended in lOOμl PBS + NHS. The cells were analysed (10,000 cells per analysis) with an Epics 541 Flow Cytometer using a 488nm laser light source and green light fluorescence detection. Figure 11 is an example of the data obtained from such an experiment and shows and increase in fluorescence bound to the cells incubated with the uPA-PAI-2 hybrid indicating that the protein binds to U937 cells. Additional experiments showed that PAI-2 itself did not bind to cells appreciably and that the binding of uPA-PAI-2 to U937 cells was competed by pro-uPA and could be inhibited by anti-uPA antibody but not by anti-PAI-2 antibody. These data demonstrated that the uPA-PAI-2 hybrid protein was able to bind to cells via the uPA receptor. Both the uPA growth factor domain and the PAI- 2 domain of the hybrid protein were thus demonstrated to be functional. Example 2

A human α₁-antitrypsin cDNA, identical in sequence to that described in the Genbank database, release 60, accession number X01683 V00496, locus HUMAIATR was amplified by Polymerase Chain Reaction using a Perkin Elmer Cetus DNA thermal cycler according to the manufacturer's instructions. The DNA sequence of the two primary oligonucleotides were:

5' Oligonucleotide

5'-GAGGATCCCCAGGGAGATGCTGCCCAGAAG-3 •

3' Oligonucleotide

5'-GGGGAAGCTTTTATTTTTGGGTGGGATTCACCACTTTTCC-3 ^■

Upon amplification both the 5' and 3 ' termini of the a - antitrypsin sequence became modified as follows:

5' Terminus

original

G GTC CCT GTC TCC CTG GCT GAG GAT CCC CAG GGA GAT GCT GCC CAG

V P V S L A E D P Q G D A A Q

I i

BamHI AAG K

modified

GAG GAT CCC CAG GGA GAT GCT GCC CAG AAG

E D P Q G D A A Q K

I I

Ba-uHI

3 ' Terminus

original GGA AAA GTG GTG AAT CCC ACC CAA AAA TAA CTGCCTCTCG G K V V N P T Q K *

modified GGA AAA GTG GTG AAT CCC ACC CAA AAA TAA AAGCTTCCCC

G K V V N P T Q K * I 1

Hindi I I

These modifications remove the 23 amino acid signal sequence and introduce a Hindlll restriction site at the 3 ' end of the cDNA.

The 1.22kbp modified cDNA was purified, digested with Hindlll and BainHI inserted and cloned into the Hindlll-BamHI sites of M13mpl9 (Yanisch-Perron et al (1985) Gene 33, 103-119) generating pDBAl (Figure 12) . The integrity of human 0^- antitrypsin was confirmed by dideoxynucleotide sequencing. A sequence encoding α^_^-antitrypsin Pittsburgh (α^AT-P) (Owen et al, 1983) was created using oligonucleotide-directed mutagenesis of pDBAl using the oligonucleotide shown below

5'-GAGGCCATACCCAGGTCTATCCCC-3 ' This resulted in a change in the codon for methionine 358 such that it coded instead for an arginine ie ATG —> AGG. This plasmid is pDBA2 (Fig 12) .

Two oligonucleotides were synthesised and then annealed to form a linker (Linker 7) to encode the C-terminal portion of the uPA growth factor domain (amino acids 24-47) and the start of the α^AT-P sequence.

Linker 7

Y F S N I H W C N C P K K F AGTAC TTC TCC AAC ATT CAC TGG TGC AAC TGC CCA AAG AAA TTC

CATGTCATG AAG AGG TTG TAA GTG ACC ACG TTG ACG GGT TTC TTT AAG

I 1

Seal

G G Q H C E I D K S E GGA GGG CAG CAC TGT GAA ATA GAT AAG TCA GAG CCT CCC GTC GTG ACA CTT TAT CTA TTC AGT CTCCTAG

I I

BamHI This linker was ligated with Kpnl-BamHI digested double stranded pDBA2 to form pDBUAl (Fig 12) .

A second linker representing the PRBl ATG environment, a methionine initiation codon and DNA encoding the N-terminal portion of the uPA growth domain (amino acids 1-23) was assembled by annealing the four oligonucleotides shown below: -

Linker δ oligo 7

M S N E L H Q V AATTAAGCTTAACCTAATTCTAACAAGCAAAG ATG AGC AAT GAA CTT CAT CAG GTA

TTCGAATTGGATTAAGATTGTTCGTTTC TAC TCG TTA CTT GAA GTA GTC CAT l i

Hindlll

I oligo 6

oligo 4

J L oligo 5 T A

Double stranded pDBUAl was digested with Seal and EcoRI and then ligated with linker 8 to form pDBUA2 (Fig 12) . The Hindlll fragment of pDBUA2 was then introduced into the expression plasmid pDBP6 to form pDBUA3 (Fig 12) . This plasmid was then introduced into a suitable S. cerevisiae strain wherein it effected production of the hybrid protein which was recognised by anti-a_xAT and anti-uPA antibodies in Western blotting experiments.

REFERENCES (incorporated herein by reference)

Appella, E. et al (1987) J. Biol . Chem . 262, 4437-4440

Beggs, J.D. (1978) Nature 275, 104-109

Beggs, J.D. (1981) Molecular Genetics in Yeast , Alfred Benzon Symposium 16, 383-395

Danø, K. et al (1985) Adv. Cancer Res . 44, 139-266

Goodey, A.R. et al (1987) In Yeast Biotechnology, 401-429, Edited by Berry, D.R., Russell, I. and Stewart, G.G. Published by Allen and Unwin

Jones et al (1986) UCLA Symp. Mol . Cell Biol . New Ser . 33, 505-518

Kirchheimer, J.C. and Remold, H.G. (19δ9a) Blood 74, 1396-1402

Kirchheimer, J.C. and Remold, H.G. (1989b) J. Immunol . 143, 2634-2639

Kruithof, E.K.O. et al (1986) J. Biol . Chem . 261, 11207-11213

Miles, L.A. and Plow, E.F. (1988) Fibrinolysis 2, 61-71

Miles, L.A. et al (1988) J . Biol . Chem . 263, 11928-11934 Moehle et al (1987a) Genetics 115, 255-263

Moehle et al (1987b) Mol . Cell . Biol . 7 , 4390-4399

Neuberger, M.S. et al (1988) 8th International Biotechnology Symposium Part 2, 792-799

Norrander, J. et al (1983) Gene 26, 101-106

Ossowski, L. (1988) Cell 52, 321-328

Ossowski, L. and Reich, E. (1983) Cell 35, 611-619

O'Sullivan et al (1979) Anal. Biochem . 100, 100-108

Owen et al (1983) W. Engl . J. Med . 309, 694-698

Pδllanen, J. et al (1990) Proc. Natl . Acad . Sci . USA 87, 2230-2234

Reich, E. et al (1987) Cancer Res . 48, 3307-3312

Saheki, T. & Holzer, H. (1975) Biochem . Biophys . Acta 384, 203-214

Sambrook, J. et al (1989) Molecular cloning : a laboratory manual, 2nd edition . Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. Tan, L.K. and Morrison, S.L. (198δ) Adv. Drug Deliv. Rev. 2 , 129-142

Thorsen, S., Philips, M. , Selmer, J. , Lecander, I. and Astedt, B. (1988) Eur. J. Biochem . 175, 33-39

Twigg, A.J. and Sherratt, D. (1980) Nature 283, 216-218

Vassalli, J.D. et al (1985) J. Cell Biol . 100, 86-92

Vu, T-K.H., Hung, D.T., Wheaton, V.I. and Coughlin, S.R. (1991) Cell 64, 1057-1068

Williams, G. (1988) riJbtec 6, 36-42

Wolf, D. and Ehmann, C. (1979) Eur. J. Biochem . 98, 375-384

Wun, T.C. and Reich, E. (1987) J. Biol . Chem . 262, 2646-3653

Yanisch-Perron et al (1985) Gene 33, 103-119

Zola, H. (1988) "Monoclonal Antibodies : a Manual of Techniques", C.R.C. Press

Zubenko, G. and Jones, E. (1981) Genetics 97, 45-64

Claims

1. A molecule comprising a first region which is capable of binding to a tumour cell and a second region which is capable of inhibiting a first protease.

2. A molecule according to Claim 1 wherein the first region binds to a receptor for a second, tumour-associated protease which may be the same as or different from the first protease.

3. A molecule according to Claim 2 wherein the second tumour- associated protease is uPA.

4. A molecule according to Claim 2 or 3 wherein the first region binds to a receptor in such a way as to block binding thereto by the second protease.

5. A molecule according to Claim 4 wherein the first region comprises a receptor-binding portion of the second protease or a receptor-binding variant of the second protease.

6. A molecule according to any one of the preceding claims wherein the first protease is thrombin.

7. A molecule according to any one of Claims 1 to 5 wherein the first protease is tumour-associated.

8. A molecule according to Claim 7 wherein the first protease is uPA, plasmin, stromelysin or cathepsin.

9. A molecule according to Claim 8 wherein the first region comprises a peptide corresponding to amino acids 20-30 of mature uPA.

10. A molecule according to Claim 9 wherein the first region comprises a peptide corresponding to amino acids 12 to 32 of mature uPA.

11. A molecule according to Claim 10 wherein the first region comprises a peptide corresponding to amino acids 1 to 47 of mature uPA.

12. A molecule according to Claim 10 or 11 wherein the first region comprises a non-plasminogen-activating mutant of uPA or pro-uPA.

13. A molecule according to any one of Claims 9 to 12 wherein the second region is PAI-2 or a uPA-inhibiting variant or fragment thereof.

14. A molecule according to any one of Claims 9 to 12 wherein the second region is α^-antitrypsin (Pittsburgh) or a uPA- inhibiting or plasmin-inhibiting variant or fragment thereof.

15. A molecule according to any one of the preceding claims comprising a polypeptide consisting of the first and second regions and, optionally, an intervening amino acid sequence combined to form a single, linear amino acid sequence.

16. A nucleotide sequence encoding a polypeptide as defined in Claim 15.

17. A process of producing a molecule according to any one of Claims 1 to 15 comprising (a) preparing the said first and second regions and joining them together or (b) preparing the first and second regions and, optionally, an intervening amino acid sequence, as a single linear polypeptide by expressing a nucleotide sequence encoding the polypeptide in a suitable host cell transformed with the nucleotide sequence.

18. A process according to Claim 17 (b) wherein the host cell is a Saccharomyces cerevisiae cell.

19. A method of combatting neoplasms comprising administering to a patient a molecule according to any one of Claims 1 to 15.

20. A compound which binds specifically to uPAR and to PAI-2.

21. A compound according to Claim 20 wherein the compound is a bispecific antibody.

22. A pharmaceutical composition comprising a molecule according to any one of Claims 1 to 15 or a compound according to Claim 20 or 21.